My experience migrating from Mach3 to UCCNC

[derek]

I’m starting this post to document the good the bad and the ugly I encounter while I convert 2 and possibly 4 machines from Mach3 to UCCNC. I’m hoping this can be a thread where people add their experiences as well. I don’t want it to become a thread that bashes the software as I’m not interested in that drama. Andy B already has a thread on macro sharing. Maybe a thread on custom button/screen sharing would be a good addition as well.

So here we go

Similar Threads:

• Honest Appraisal of UCCNC
• Mach 3 vs. UCCNC- any thoughts / experience ?
• Migrating to Sprutcam 9
• Newbie- migrating/upgrading Alphacam to new PC
• reports folder not migrating

[derek]

First off Balazs recommends 1.6GHz or higher. Heed that. I tried to run it on a AMD sempron 1.5GHz and it was ugly. Lots of lag, slow startup and it would fault the drives during G0 moves. I moved it over to a Pentium4 dual core with 2.8 GHz and XP pro and it ran great. Night and day difference. So from here on out those specs are going to be my baseline for the other machines.

On any screen that has “Apply” and “Save” make sure you use them. It looks like there may be an issue where if you make a change and leave the screen with out saving not only does it not save the change but it doesn’t reverse what you did. So you go back to the screen to see why “Softlimits” doesn’t work, It’s still checked but because you didn’t hit apply and save it never took.

I used the Mach3 XML importer and it worked great. I had to uncheck active low on step and direction as the travel was backwards. X,Y,and Z axis all showed to be slaved to A so I changed that to none. Remember apply save. It’s you’re friend

At that point everything was operational. My homing switches, probe input and spindle worked fine. The spindle has manual speed control so I don’t know how well the converter handles that part of the spindle speed settings.

All screens will let you jog the machine but when you are in the configuration section there seems to be a lag and you may jog farther than you intended.

There are no tool flyout button descriptions so resist the “what’s this button do?” urge
P1 P2 and P3 run macros that will blow right past you’re soft limits without a care in the world. I think they are for parking the machine.

The MDI is a little different than Mach. The only way to access it is though a mouse click in the field. It’s also not as apparent that it’s active as Mach. I’m getting used to it but I would love it if hitting the enter key made it active like Mach. The up and down arrow keys let you move through prior MDI commands.

If you use fixed length tool holders tool length management blows. There is no button to set the tool length to the corresponding tool number in the tool table. I must be done manually. I’m conversing with Balazs on this but it’s probably going to be low on his list. Also only having 20 tools in the table is going to require I change how I handle tool numbers in my cam program. Hopefully that will be expanded at some point.

Well that’s about as far as I’ve gotten with it. I’ll be using it in anger next week so we’ll see how it goes. I’ll continue to add any quirks/differences from mach that I run into.

Keep in mind Mach got the features it has because users participated and made it happen.

[derek]

Set my first Hot Keys. I wanted the spacebar to turn on Feed Hold. Navigated to Configuration-IO Trigger-hotkeys. The document "Buttons By Number" in the documents folder has all the button codes for the functions. pushing the set button on "key code" allows you to push a keyboard key and pushing set in the function fields brings up a list of available options. There is no way to search so I found that locating the number in the documentation was quicker. I also set "control" for incremental jog and "shift" to return to continuos. I had to input the keyboard function numbers manually as it wouldn't accept shift or control with the set button.

Derek

[derek]

So here are some results from my first real run.

I have a large gantry router with a lot of weight on it. Although the servos can move it at 120 IPM the gantry isn’t rigid enough to reasonably stop it at that speed. So it’s always been a compromise between speed, accuracy, and mechanical overshoot. Even in the best cases I got bad chatter marks in a lot of the tight corners. Mach has never been great at CV and is really quite bad. Over time I settled into settings that would give me a reasonable cutting speed and still avoid big gouges when Mach tried to change direction too fast. I had my servos set with 120 ipm max with 12 acceleration. Any more acceleration and it would really shake the gantry during direction changes. Any less and accuracy would really suffer. My cutting speeds were generally in the 60-80 ipm . To get a really good cut with no overshoot chop I would need to run it at 30 ipm. The parts I make are 3D carvings (foundry patterns) and take hours to carve. Because of the nature of my parts I tend to use a horizontal cutting strategy.

So I decided to use the CV settings Andy B recommended in his youtube video. I ran some known code and it was noticeably better. Still a little choppy but better. I increased the feed from 60 to 100 and that’s where you could really see the planner working. It would run at 100 ipm on a straight and slow down to 40 or so in the corners and then accelerate back to 100. The acceleration at 12 was still shaking the gantry a tiny bit. I decided to try and reduce my acceleration on the X and Y to 6. This made a huge difference in the gantry shake. It also changed the speed of entry into the corners which showed me the planner was doing it’s job.

I did a quick accuracy test. I cut a slot .5” wide with a .25” radius at the end. It was an open slot and I entered it at 100 ipm. I used a .25” cutter and did it in multiple passes. I used a radius gauge to check the arc at the end of the slot. It was pretty close. and close enough for what I was doing. I did the same test at an acceleration of 12 and the results were identical. I didn’t have time for any other testing as I had a pattern to carve.

So now the results. I’m carving a pattern that’s a inline 6 valve cover. The roughing passes went great. A little longer of course but the quality was really good. I then did the finish work. My cam program gave me a run time of 13 hours 17 min. I know from experience this is pretty accurate. I ran it all night last night and it was just finishing up when I got here this morning. It finished at 15 hours and 14 min. So all the slowing down added 2 hours or 16 percent to the cut time. The 100 ipm cuts helped a lot because of the long straights of the valve cover. But it was the best looking production piece I’ve ever carved with this machine. Zero chop in the corners. So for me the extra time was worth it.

[ger21]

If you have a lot of long straight moves, then you should increase the feedrate even higher. It might cancel out the time lost on the slower acceleration.

Good to hear more good things about the CV quality.

[Signal9]

Looks like latest version has tool lengths, but diameter of tool is not functional yet. It also appears to be more than 20 tools now too.